Larry Hammick
unread,Aug 21, 2009, 10:57:27 PM8/21/09Sign in to reply to author
Sign in to forward
You do not have permission to delete messages in this group
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to Open Mathematics Forum
Given a positive integer and a square matrix of this form
a Db
b a
and another matrix of the same form
x Dy
y x
the product of those two has the same form, voila
ax+Dby .. D(bx+ay)
bx+ay .. ax+Dby
Also, the product of the two given determinants is the determinant of
the product of the two given matrices. Leaving some details to the
reader, we thus get (by transport of structure) a group structure on
the set of lattice points on the parabola
(1) xx - Dyy = 1
or (same thing) a group structure on the set of solutions of Pell's
equation.
The traditional account of Pell's equation uses the ring of real
numbers of the form u+sqrt(D)v where u and v are integers, but the
matrices avoid these unsightly square roots and, at the same time,
they impart an attractive geometric tone to the subject.
Consider now the set of solutions of (1) having x>0. If D is
squarefree it contains more than just the identity element (1,0) of
the aforementioned group. The set of solution-matrices is generated by
the single element for which y>0 and x is as small as possible, and if
we fix D=2 that generator is
3 4
2 3
The first few powers of this matrix give the first few solutions (x,y)
of (1), namely
(1,0) (3,2) (17,12) (99,70) (577,408)
Notice that the these triangles:
(0,0) (1,0) (3,2)
(0,0) (3,2) (17,12)
(0,0) (17,12) (99,70)
(0,0) (99,70) (577,408)
etc.
all have the same area, 1.
And I think I'll leave it at that, for anybody who wants to pursue
this little recreational topic.